Older adults are disproportionately affected by musculoskeletal pain conditions and receive more opioid analgesics than younger persons, but benefits and harms of opioid therapy for persistent musculoskeletal pain in older adults are unclear. The objectives of this study are to evaluate potential adverse effects of opioid analgesics on falls, fractures, and functional outcomes in older men with persistent back, hip, or knee pain and to examine whether opioid-associated risk varies according to the presence of key geriatric risk factors (i.e., baseline functional impairment, cognitive impairment, and poor physical performance). In a cohort of older men with persistent back, hip, or knee pain, the study aims to evaluate the following: 1) associations between opioid analgesic use and risk of incident and recurrent falls; 2) associations between opioid analgesic use and risk of incident clinical fractures; and 3) associations between opioid analgesic use and decline in self-reported physical function and objective measures of physical performance. This study will analyze existing longitudinal data collected in an NIH-funded prospective cohort study of community-dwelling men age 65 years and older, the Osteoporotic Fractures in Men (MrOS) study. MrOS participants completed a baseline examination between 2000 and 2002 and two subsequent examinations, approximately 4.5 years and 7 years later. Current medications, dose, and frequency of use were recorded from medication containers that participants brought to each examination. Assessments at each examination included measures of objective physical performance (including strength, dynamic balance, and gait speed), self-reported physical function (including limitations in independent activities of daily living and physical tasks), cognitive function, and self-reported physical activity. Participants were contacted every four months between visits to determine occurrence of new falls or fractures. For this study, the analytic cohort includes MrOS participants who reported back, hip, or knee pain most or all of the time at baseline. Participants who reported opioid analgesic use will be compared with those who used only non-opioid analgesics and those who reported no analgesic use. Propensity scores will be used to control for baseline differences between analgesic use groups. Generalized estimating equations will be used to model the relationship between opioid use and the primary outcome of incident falls (aim 1). As a secondary measure, the outcome of recurrent falls will be evaluated using logistic regression. Associations between opioid use and incident clinical fractures will be examined using survival analysis to model time to first fracture (aim 2) Geriatric risk factors and their interactions with analgesic category will be included as additiona covariates in subsequent models to evaluate their effects on associations between opioids and falls or fractures. Logistic regression will be used to model the relationship between opioid use and decline in a) physical function and b) physical performance (aim 3).